Stabilisation of the Foot and Ankle Complex
Proactive and Reactive Responses to Disturbances in the Frontal Plane
©2005
Doktorarbeit / Dissertation
116 Seiten
Zusammenfassung
Inhaltsangabe:Abstract:
Probably one of the main contributions of this thesis has been the use of a three-dimensional kinematic model accounting not only for ankle motion but also for the motion of the lateral and medial columns of the forefoot with regard to the rearfoot (Arampatzis et al., 2002), in a joint stability context. The obtained values may serve as reference for the planning of further studies and provide a base for building up new hypotheses. However this thesis did not aim to merely describe the kinematics but rather to provide more knowledge regarding the stabilisation of the foot and the ankle. Therefore another important contribution is surely the simultaneous study of the kinematics, the EMG and the ground reaction forces, which allows a better understanding of the whole stabilisation process.
The presented results have shown that forefoot motion is fundamental in foot and ankle stabilisation. The flexibility of the forefoot, especially in the frontal plane, permits a fast and appropriate adaptation to the ground. Furthermore the high mobility of the forefoot, allows the ankle to rotate slower and to a lesser extent. Possibly this reduction in required ankle motion can contribute considerably to injury prevention, since the forces acting at the ankle are high and a misalignment with regard to the ground reaction forces could rapidly lead to moments overwhelming the stabilising potential of the involved structures. In addition, the rapid adaptation of the forefoot to the ground can potentially provide more precise and earlier feedback regarding the ground characteristics than the structures surrounding the ankle joint. This way the corresponding adjustments in an immediate feedback could happen earlier, and the consequences of future interactions could be predicted more accurately.
The results from the presented studies support the notion that joint stabilisation does not rely primarily on proprioception. Prolonged peroneal latencies might in fact be due to deafferentiation consequent to the recurrent sprains. However prolonged latencies do not seem to be responsible for a functional instability. On one hand the differences in latency times between healthy and unstable ankles are relatively low and not consistently observed. Those studies identifying prolonged latencies in functionally unstable joints, report differences close to 15 ms (Konradsen and Ravn, 1990; Löfvenberg et al., 1995). Fifteen ms is a short time to have a high […]
Probably one of the main contributions of this thesis has been the use of a three-dimensional kinematic model accounting not only for ankle motion but also for the motion of the lateral and medial columns of the forefoot with regard to the rearfoot (Arampatzis et al., 2002), in a joint stability context. The obtained values may serve as reference for the planning of further studies and provide a base for building up new hypotheses. However this thesis did not aim to merely describe the kinematics but rather to provide more knowledge regarding the stabilisation of the foot and the ankle. Therefore another important contribution is surely the simultaneous study of the kinematics, the EMG and the ground reaction forces, which allows a better understanding of the whole stabilisation process.
The presented results have shown that forefoot motion is fundamental in foot and ankle stabilisation. The flexibility of the forefoot, especially in the frontal plane, permits a fast and appropriate adaptation to the ground. Furthermore the high mobility of the forefoot, allows the ankle to rotate slower and to a lesser extent. Possibly this reduction in required ankle motion can contribute considerably to injury prevention, since the forces acting at the ankle are high and a misalignment with regard to the ground reaction forces could rapidly lead to moments overwhelming the stabilising potential of the involved structures. In addition, the rapid adaptation of the forefoot to the ground can potentially provide more precise and earlier feedback regarding the ground characteristics than the structures surrounding the ankle joint. This way the corresponding adjustments in an immediate feedback could happen earlier, and the consequences of future interactions could be predicted more accurately.
The results from the presented studies support the notion that joint stabilisation does not rely primarily on proprioception. Prolonged peroneal latencies might in fact be due to deafferentiation consequent to the recurrent sprains. However prolonged latencies do not seem to be responsible for a functional instability. On one hand the differences in latency times between healthy and unstable ankles are relatively low and not consistently observed. Those studies identifying prolonged latencies in functionally unstable joints, report differences close to 15 ms (Konradsen and Ravn, 1990; Löfvenberg et al., 1995). Fifteen ms is a short time to have a high […]
Leseprobe
Inhaltsverzeichnis
ID 8835
Morey-Klapsing, Gaspar Maximilian Gabriel: Stabilisation of the Foot and Ankle Complex
- Proactive and Reactive Responses to Disturbances in the Frontal Plane
Hamburg: Diplomica GmbH, 2005
Zugl.: Deutsche Sporthochschule Köln, Dissertation / Doktorarbeit, 2005
Dieses Werk ist urheberrechtlich geschützt. Die dadurch begründeten Rechte,
insbesondere die der Übersetzung, des Nachdrucks, des Vortrags, der Entnahme von
Abbildungen und Tabellen, der Funksendung, der Mikroverfilmung oder der
Vervielfältigung auf anderen Wegen und der Speicherung in Datenverarbeitungsanlagen,
bleiben, auch bei nur auszugsweiser Verwertung, vorbehalten. Eine Vervielfältigung
dieses Werkes oder von Teilen dieses Werkes ist auch im Einzelfall nur in den Grenzen
der gesetzlichen Bestimmungen des Urheberrechtsgesetzes der Bundesrepublik
Deutschland in der jeweils geltenden Fassung zulässig. Sie ist grundsätzlich
vergütungspflichtig. Zuwiderhandlungen unterliegen den Strafbestimmungen des
Urheberrechtes.
Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in
diesem Werk berechtigt auch ohne besondere Kennzeichnung nicht zu der Annahme,
dass solche Namen im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei
zu betrachten wären und daher von jedermann benutzt werden dürften.
Die Informationen in diesem Werk wurden mit Sorgfalt erarbeitet. Dennoch können
Fehler nicht vollständig ausgeschlossen werden, und die Diplomarbeiten Agentur, die
Autoren oder Übersetzer übernehmen keine juristische Verantwortung oder irgendeine
Haftung für evtl. verbliebene fehlerhafte Angaben und deren Folgen.
Diplomica GmbH
http://www.diplom.de, Hamburg 2005
Printed in Germany
Gaspar Maximilian Gabriel Morey Klapsing
was born in Essen (Germany) in 1971.
In 1978 he moved to Palma de Mallorca (Spain) with his family.
In 1989 he moved to Valencia (Spain) where he studied physical education at the University
of Valencia (Institut Valencià d'educació física).
From February 1994 to February 1995 he enjoyed a grant from the Polytechnic University of
Valencia at the IBV (Institut de Biomecànica de València) in the group of sport biomechanics.
In 1994 he got his diploma in physical education (Licenciatura en educación física) and
in 1995 the certificate for teaching aptitude (CAP: Certificat d'aptitud pedagògica). This same
year he visited a course in neurophysiological techniques in clinical practice at the University
of the Balearic Islands. Short time later he moved to Cologne (Germany) where he joined the
research group led by Prof. Brüggemann at the German Sport University Cologne and
accomplished his doctorate studies.
In 1997 he became scientific assistant at the Institute for Athletics and Gymnastics.
In 2001 the research group moved to the institute for Biomechanics (now Institute for
Biomechanics and Orthopaedics). Later this year he became scientific co-worker of the
institute. In the last years he participated in several research projects mainly related to
footwear, sport materials and movement analysis. His main interests in basic research are in
the fields of joint stabilisation, muscle mechanics and more recently motor control as well. He
authored and co-authored more than 50 papers and abstracts in international peer reviewed
journals and congress proceedings. In addition he translated a book on infantile gymnastics
and several technical texts (e.g. DIN norms on sport surfaces) from the German language into
Spanish.
Abstract
This thesis comprises four studies all committed to a better understanding on how
functional joint stabilisation is achieved. In order to do so the neuromechanical behaviour
of the foot and ankle complex was experimentally studied under different conditions. The
first study focused on EMG onset times in response to sudden expected and unexpected
tilts, as onset times have been often studied in relation to functional ankle instability and
the literature reports controversial results. We found onset time determination to have
several inherent methodological problems. In addition their relevance with regard to joint
stability remained unclear. Consequently onset times were not considered in the following
three studies.
The methods were similar among the remaining studies. The kinematics were determined
by means of a three dimensional model of the foot and ankle comprising 4 segments
(shank, rearfoot, lateral forefoot column and medial forefoot column). This allowed a more
detailed and functional analysis of foot motion, which is an important issue because
observing only one of the foot joints does not allow to predict the behaviour of the others.
The electromyographic (EMG) signals from six leg muscles as well as the three
dimensional ground reaction forces (GRF) were also analysed.
The second study, where expected and unexpected sudden tilts were compared, showed
that expectedness could improve joint stabilisation in that the same kinematics and lower
GRF were achieved with lower EMG amplitudes. As no differences were observed
immediately prior to plate release, it can be stated that the enhanced stabilisation
observed for expected trials had to arise from supraspinal influences.
The third study observed sudden unexpected tilts in lateral or medial direction (direction
was known). Direction specific as well as common responses to the tilts could be
observed in most parameters. Higher mediolateteral ground reaction forces and reduced
kinematics with no general increase in muscular activation were observed for the medial
tilts. This suggests, that passive structures may counteract destabilising forces and this
way reduce the otherwise needed muscular activation.
The fourth study dealt with a more functional task: Landings from 40 cm height onto three
surfaces with different frontal plane inclination. Contrary to the previous studies, where it
was tried to keep the initial conditions identical until plate release, this time also
adjustments made prior to the perturbation (ground contact) could be analysed. The
results revealed specific adjustments in foot motion and EMG prior to TD which were
necessarily planned before touchdown. Early specific post-TD responses not related to
any feedback arising from the collision with the surface were also observed. It is
suggested that these are mainly due to self-stabilizing mechanisms of the
musculoskeletal system. It was concluded, that different central motor commands were
produced in response to the different surfaces. It was hypothesised that the aim of these
adjustments is to enhance the self-stabilising potential of the involved structures.
Analysing the results from all four studies together, evidence arises, that stabilisation is
mainly achieved by mechanisms others than direct proprioceptive feedback of the acute
perturbation. These mechanisms involve prediction of the perturbation to come and
passive mechanics.
Acknowledgements
From the academic point of view, which for the last years I could not really
separate from my private life, and in chronological order, I would like to
thank the people at the IBV (Institut de Biomecànica de València) where I
really enjoyed my first steps in the field of biomechanical research. Ilona
Gerling built the bridge for me to come over to the German Sport
University of Cologne. Prof. Brüggemann opened the doors of his lab to
me and shared his enthusiasm on biomechanics and a few of the millions
of ideas crossing his head. Dr. Arampatzis tutored me and was my
constant scientific guide and conscience during my research activities.
And the colleagues? Of course, the colleagues. Especial thanks go to Dr.
Uwe Kersting and Dr.Toni Arndt who took me by the hand during my first
steps in Cologne (not only in the lab). Dr. Mark Walsh shared countless
hours sitting on the same desk with me ("mucho trabajo, poco dinero").
Gianpiero De Monte was a great help during my measurements and for
calming the yearn for mediterranean communication ("tante parole che qui
non posso scrivere"). Especial thanks go also to Lida Mademli for
checking, re-checking, re-re-checking ... the whole text for typos, format
congruence, etc.
Finally I want to thank everyone supporting me during these years: Thank
you for your caring help and for being there.
Dedication
To my parents...
... who always let me go and do whatever I thought to be right,
supported me in doing it and always kept their arms open for me to
come back or hold me when ever it was necessary.
CONTENTS
PROLOGUE ... 1
INTRODUCTION AND OUTLINE... 2
1
First Study: Onset times and joint stabilisation ... 7
1.1
INTRODUCTION ... 10
1.2
METHODS... 12
1.3
RESULTS ... 15
1.4
DISCUSSION ... 19
1.5
REFERENCES ... 21
2
Second Study: Joint stabilising response to expected and
unexpected tilts ... 24
2.1
INTRODUCTION ... 28
2.2
METHODS... 30
2.3
RESULTS ... 37
2.4
DISCUSSION ... 41
2.5
CONCLUSIONS ... 45
2.6
REFERENCES ... 45
3
Third Study: Joint stabilising response to lateral and medial
tilts ... 48
3.1
INTRODUCTION ... 51
3.2
METHODS... 53
3.3
RESULTS ... 59
3.4
DISCUSSION ... 63
3.5
CONCLUSIONS ... 66
3.6
REFERENCES ... 67
4
Fourth Study: Foot and ankle stabilisation during drop
landing: A kinematic, kinetic and electromyographic study .. 69
4.1
INTRODUCTION ... 72
4.2
METHODS... 75
4.3
RESULTS ... 81
4.4
DISCUSSION ... 87
4.5
REFERENCES ... 93
SUMMARY AND CONCLUSIONS ... 97
BIBLIOGRAPHY ... 100
1
PROLOGUE
At the very beginning of my PhD studies, I was supposed to investigate
how the muscles of the lower leg control landing and jumping tasks and
how this control could influence the storage and delivery of mechanical
energy in the foot structures. Special attention should be paid to joint
stability and to frontal plane foot motion. One of my first duties was to find
out how to classify stable and unstable subjects. Therefore I needed a
definition of joint stability. After intensive searching through the literature
and consulting experts, I realised that there was no definition suited to my
purpose. The nature and the factors affecting joint stability are too
manifold. Any objective grouping criterion would describe only a minor
aspect of joint stabilisation. The relationship between functional joint
stability and classical clinical tests aiming to quantify the degree of
mechanical stability (e.g. stress radiography) proved to be low. Finally I
had read and thought a lot about the function and the stabilisation of the
foot and the ankle complex. It became clear to me that the issue of joint
stabilisation is far away from being well understood. So at the end it was
decided to try to provide some more insight into this phenomenon. As the
famous German tale writer Michael Ende would have said: The issue on
the storage and delivery of mechanical energy in the foot and ankle
complex is now "another story and shall be told another time".
In the following, four studies concerned with foot and ankle stabilisation
are exposed. As each study is intended to be a stand-alone report, many
of the methods, the exposed ideas and conclusions are repeated.
However each study provides new knowledge and enlightens the issue of
joint stabilisation of the foot and ankle complex a bit further. The first three
studies are published or in press. You will find some differences among
studies in the nomenclature albeit describing the same thing. This is due
to the different referees reviewing each study, who exhorted me to do so.
The articles are presented as they are published except for the formatting
which has been unified throughout the thesis.
Introduction and outline
2
INTRODUCTION AND OUTLINE
This thesis focuses on the foot and ankle complex and its stabilisation.
There is a reasonable amount of studies done under passive conditions
using anatomical preparations (Cass et al., 1984, Siegler et al., 1994) or
done with living subjects (Seligson et al., 1980; Liu et al., 2000). These
studies provided information regarding the mechanical constraints of foot
motion and concentrated almost exclusively on the tibio-talar and the talo-
crural joints. Another considerable amount of studies oriented their efforts
to elucidate the effectiveness of different orthoses or tape in preventing
excessive ankle motion or injuries and to their possible interference with
performance (Karlsson and Andreasson, 1992; Eils and Rosenbaum,
2003; Leanderson et al., 1999). In these studies the kinematics were often
disregarded or analysed with very simple two-dimensional models.
Furthermore again attention was paid only to the tibio-talar and the talo-
crural joints, whereas forefoot motion has not been considered. In general
the effectiveness of orthoses, at least in reducing the amount of recidives,
is supported (Verhagen et al., 2000; Tropp et al., 1985). Other studies
tried to identify factors associated to functional ankle instability, i.e. factors
favouring recurrent sprains (Neely, 1998; Karlsson et al., 1997). Some of
the studied factors have been: a) Mechanical instability, which has proven
to be associated to functional instability. However the correlation is weak
and it is not possible to estimate the degree of the one from the other
(Tropp et al., 1985; Vaes et al., 2001). b) Latency times, especially of the
peroneal muscles. The results from the literature are controversial and
somewhat more recently their clinical relevance regarding joint
stabilisation has been also questioned (Benesch et al., 2000; Morey et al.,
2004). c) Joint position sense which seems to be impaired in recurrently
sprained ankles (Konradsen et al., 1998; Konradsen, 2002) and d) Muscle
strength. The results from the literature are differing and not conclusive
(Kaminski and Hartsell, 2002; Konradsen et al., 1998).
In contrast to the relatively high amount of literature related to factors
associated to functional instability, there is a lack of articles focusing on
Introduction and outline
3
the understanding on how stabilisation is achieved. One main shortcoming
of the existing literature is the use of too simple foot and ankle models,
mainly considering the foot as one single segment and observing only one
plane of motion. Two studies stressing the importance of forefoot motion,
especially in the frontal plane are those from Stacoff et al. (2000) and
Arampatzis et al. (2002). One further limitation of most studies in this field
is the segmented approach contemplating either neural phenomena or
mechanical ones, but not integrating them (Caster and Bates, 1995;
Nielsen, 2004). As all factors are interdependent, when observing only one
parameter and trying to provide an explanation of the results, the need to
rely on assumptions is relatively high. This is because the other involved
parameters are not known and their behaviour can at best be only
indirectly inferred.
Basing on this background, the present thesis tries to enlighten the topic of
joint stabilisation. In order to do this, different stabilisation tasks are
studied by measuring their three-dimensional kinematics, not only
between rearfoot and the lower leg, but also between the lateral and
medial forefoot columns and the rearfoot. This is done by means of a
three-dimensional multi-body system model of the shank and the foot
(Arampatzis et al., 2002). In addition the activation patterns of six muscles
of the lower leg (mm. peroneus longus and brevis, m. tibialis anterior and
the three heads of the triceps surae: gastrocnemius lateralis,
gastrocnemius medialis and soleus) are estimated using surface
electromyography (EMG) and also the three-dimensional ground reaction
forces are measured and analysed. Rather than trying to identify factors
related to joint instability, this thesis aims to improve our understanding on
how joint stabilisation is achieved.
The thesis comprises four studies. Three of them utilise sudden tilt tests in
their experimental protocol, whereas the last one focuses on landings. The
first study is a critical analysis of the use of EMG onset times with regard
to joint stabilisation studies, as this is one of the most utilised approaches.
In the three following studies different stabilising demands are
experimentally induced. The stabilising responses are then analysed with
Introduction and outline
4
regard to: the kinematics of the foot and ankle complex, the corresponding
EMG signals and the ground reaction forces.
More concisely, in the first study the issue of EMG onset times is critically
discussed in their relation to joint stability based on experimental data
obtained during tilt plate tests, which represent a typical approach in the
field of joint stabilisation studies. Several methodological concerns in their
determination are exposed. Furthermore onset times calculated using
different algorithms are compared to integrals of the EMG signal regarding
their robustness. The information provided by both parameters is
discussed. In general the onset times showed a considerably lower
repeatability than EMG integrals. In some cases earlier onset times
corresponded to lower EMG integrals and in others constant onsets to
variable integrals. It is concluded that in many cases onset times alone are
not sufficient for describing early muscle activation and when not aware of
the limitations, such studies might even induce to misleading conclusions.
The additional calculation of amplitude related EMG parameters can
provide relevant information regarding the quality of this early activation
period.
In the second study the focus is set on the influence of awareness on the
response to a sudden inverting tilt. It was assumed that awareness of the
instant of tilt would enhance joint stabilisation. The aim was to observe
how this advantage is reflected in the considered parameters (kinematics,
EMG and ground reaction forces). The kinematics themselves are of
interest, since in the literature the information on the kinematics during tilt
tests is almost completely restricted to the motion between rearfoot and
tibia. At the presented tilt test studies the kinematics are analysed only in
the eversion inversion motion of the three modelled joints. This is enough
for the purpose of the experiments because the experimental treatment
(perturbation) is in this plane of motion and therefore the main effects
should also appear in this plane. Whereas unexpected and expected trials
did not show significant differences in the kinematics, higher EMG
amplitudes and horizontal force amplitudes were found for the unexpected
trials. Opposite to that, whereas no differences in electromyographic or
ground reaction force parameters were found between stable and unstable
Introduction and outline
5
subjects (subjective feeling of stability), the kinematics revealed higher
amplitudes and velocities for the stable group. It was concluded that
awareness can enhance joint stabilisation. The results provide evidence
on that this enhancement is triggered at supraspinal levels, as there were
no significant differences in any studied parameter prior to plate release,
and the only difference between the experimental conditions was the
awareness of the instant of tilt. One further conclusion of this study is that
higher rather than earlier activation seems to be decisive in joint
stabilisation.
In the third study inverting tilts are compared to everting ones. The
followed philosophy is similar to that in the second study. This time the
perturbations are identical in magnitude but opposite in direction. This
protocol aimed to produce generic and specific responses to sudden
perturbations of joint position. These should be analysed in all considered
parameters (kinematics, EMG and ground reaction forces). Forefoot to
rearfoot motion was found to be faster and have greater amplitudes than
ankle motion. In general medial tilts showed lower motion amplitudes and
angular velocities than lateral tilts but higher horizontal ground reaction
force integrals. Interestingly, despite of opposite tilt directions, the EMG
patterns were similar for both conditions, indicating that the temporal
characteristics of the EMG triggered by joint position perturbations
correspond to generic responses which are not, or only weakly related to
the direction of the perturbation. The EMG responses showed also some
direction specific differences in the amplitudes. The higher mediolateral
ground reaction forces, together with the reduced kinematics and no
general increase in muscular activation during medial tilts suggest that in
this direction the contribution of passive structures to counteract the
destabilising forces is higher and sufficient to reduce the otherwise needed
muscular activation. This provides evidence on that the contribution of
passive structures to joint stabilisation can vary depending on the
geometry of the joints and the destabilising forces.
The fourth study, which deals with foot and ankle stabilisation after
landings, aimed to examine a more functional task than sudden tilts during
quiet standing. Landings are a good model for the study of joint stabilisa-
Introduction and outline
6
tion (Duncan and McDonagh, 2000; Pelland and McKinley, 2004).
Whereas at the former studies of this thesis using tilt plate tests it was
tried to keep the initial conditions identical until the instant of the
perturbation, at this study the initiation of the fall is controlled, but during
the fall there is time to prepare the collision with the ground. This allows
changing the orientation of the segments and the activation of the muscles
prior to the perturbation (collision). Similar to both former studies the
responses of the considered parameters to three different surface
conditions (level surface or 3° inclined either medially or laterally) are
analysed and compared. Surface specific responses were observed in the
kinematics and in the EMG even prior to touchdown, e.g. higher lateral
forefoot inversion and peroneal activity for trials onto the laterally inclined
surface. The specificity of the response was higher for both forefoot joints
than for the ankle joint, especially in eversion-inversion. Similarily the
peroneal muscles were more sensitive to surface inclination than the
muscles of the triceps surae. The medially inclined surface led to lower
mediolateral ground reaction forces near touchdown, and to a lower
vertical force maximum than the laterally inclined surface. All these results
indicate that early post-touchdown responses can be explained by self-
stabilising mechanisms of the musculoskeletal system which are not
related to any feedback arising from the collision with the ground. From
this study it is concluded, that changes in surface condition can produce
different central motor commands. It is suggested that these motor
commands aim to enhance the self-stabilising potential of the whole
system. This way the need to rely on the relatively slow feedback
mechanisms is reduced and the neuromuscular system is relieved. Finally
this interpretation of the data, provides a frame which allows explaining the
mechanisms behind the success of proprioceptive training in reducing
recurrent injuries (Tropp et al., 1985 and 1988; Lephart et al., 1997;
Verhagen et al., 2000), without the need to assume that joint stabilisation
is largely dependent on proprioception.
Chapter one
Onset times and joint stabilisation
7
1 First Study: Onset times and joint stabilisation
In most of our joints, an adequate muscle activity is necessary to maintain
stability. So muscle activity is one of the important factors to account for
when studying joint stabilisation. A considerable amount of research has
been done on the field of functional joint stabilisation. Cohen and Cohen
(1956) proposed the `arthrokinetik reflex' as a joint stabilising mechanism.
Somewhat later Freeman (1965) stated "functional instability is usually in first
place due to incoordination consequent to deafferentiation". Proprioception
became a main focus of attention in joint stability studies. In this context,
muscle onset times in response to sudden perturbations were considered to
be an indicator of proprioceptive performance (Konradsen and Ravn, 1990;
Löfvenberg et al., 1995). During the analysis of the literature concerned with
joint stability many articles focusing on EMG onset times were found.
However, our experience at the institute was that onset times were quite
variable. In contrast the amplitude or frequency related parameters were
usually far more reliable, even during the pre-innervation phase. This
motivated the following study, where several EMG onset detection algorithms
were tested and compared to a commonly utilised EMG amplitude related
parameter (the integrated EMG).
Chapter one
Onset times and joint stabilisation
8
CHOOSING EMG PARAMETERS:
Comparison of different onset determination
algorithms and EMG integrals in a joint stability study
Gaspar Morey-Klapsing
Adamantios Arampatzis
Gert Peter Brüggemann
Institute for Biomechanics, German Sport University Cologne, Cologne,
Germany
Published in:
Clinical Biomechanics (Bristol, Avon). 2004 Feb;19(2):196-201
doi:10.1016/j.clinbiomech.2003.10.010
Chapter one
Onset times and joint stabilisation
9
Abstract
Objective. The aim was to test various algorithms for onset determination and compare
onset repeatability to that from integrals of the EMG signal. The information contained in
both parameters is discussed.
Design. Onset times were calculated using six different algorithms. The integral of the EMG
signal was calculated for seven intervals: From tilt start and from each of the resulting onsets
to 200 ms after tilt start.
Background. EMG onset times are often utilised, especially regarding co-ordination patterns
or joint stability. There are almost as many different procedures for onset determination as
authors dealing with it. Results in the literature are contradictory. The determination and
usage of onset times remains controversial.
Methods. EMG signals from 6 muscles of the lower leg of 23 subjects were recorded during
three consecutive, expected and unexpected sudden inversion and eversion trials while
standing on a tilting platform.
Results. In most cases the repeatability of the onset times was considerably lower than that
of the integrals of the EMG for all studied algorithms. In some cases earlier onset times
corresponded to lower integral values and constant onsets to variable integrals.
Conclusions. In many cases onset times alone are not sufficient for describing onset
phenomena. The additional calculation of the integrated EMG might provide relevant
information regarding the quality of early activation.
Relevance
The findings are evidencing that care should be taken when interpreting onset times alone.
The additional use of the integral of the EMG signal is suggested to provide more meaningful
information.
Chapter one
Onset times and joint stabilisation
10
1.1 INTRODUCTION
Muscle onset times are often utilised in studies concerned with electromyo-
graphy (EMG), most of them regarding co-ordination patterns or joint stability.
In the literature we find pure visual determination protocols (Ebig et al., 1997
and Hodges and Bui, 1996), computer aided protocols at which the
experimenter has to decide whether or not to accept the detected onset or
where to finally place it (Hodges and Bui, 1996 and Di Fabio, 1997), as well
as fully automated algorithms (see table 1.1). Different filters and onset
thresholds are used and as stated by Hodges and Bui (1996) in many cases
the criteria or methods for onset determination are not even specified.
Sometimes a fixed value is used as threshold (Zhou et al.,1995), but in most
algorithms the threshold is based on the standard deviation or a percentage
of the EMG signal while the corresponding muscle is relaxed. As indicated by
Winter (1984), the chosen threshold strongly influences the instant of onset
detection. Further, the quality of onset time detection strongly depends on the
quality of the signal. The fact of having a broad variety of methods (Table
1.1) and the lack of consensus might indicate that none of the known
techniques is really satisfactory nor generally accepted. Benesch et al.
(2000) despite stating that peroneal reaction time is a stable and reliable
parameter, further state that its clinical relevance is not yet clear. So the
determination and usage of onset times remains a controversial topic. It is
very difficult to differentiate the real variability of a parameter from the
variability due to flaws in the detection algorithm (Tomberg et al., 1991).
However, making some assumptions, it is possible to estimate the contribu-
tion of different sources to the total variability (Brinckmann et al., 2002).
Bonato et al. (1988) developed a statistical method using a double threshold
detector. This method yields standard deviations below 15 ms when
analysing gait patterns. This is enough for many applications and may reflect
a real variability. However it is still too much for many other applications, e.g.
studies dealing with EMG latency times in response to sudden tilts, where the
reported differences between groups are often in the range or below this
standard deviation.
Chapter one
Onset times and joint stabilisation
11
The integrated EMG (IEMG) is frequently used for describing EMG activity.
There is little, if any, controversy about its use. Further, many studies have
dealt with its reliability (Gollhofer et al., 1990; Mero and Komi, 1986; Goodwin
et al., 1999; Yang and Winter, 1983). Most of them reporting high correlation
values for repeated measures.
Table 1.1.
Several authors and the corresponding methods for onset determination used.
Author Year
Onset
determination
Tomber et al.
1991 Electronic determination by a manually adjusted threshold
McKinley & Pedotti
1992 At above 95% confidence interval for baseline during more than 10 ms
Johnson & Johnson 1993 At 200% above noise
Zhou et al.
1995 At 0.015 mV above baseline value
Lynch et al.
1996
10 SD above lowest rest level in several 100 ms RMS of 200 ms resting
EMG
Ebig et al.
1997 Visual determination
Bonato et al.
1998 A statistical method using a double threshold detector
Duncan &
McDonagh
2000 Average of the ten highest data points in a fixed window (35-80 ms)
Vaes et al.
2001 At amplitude twice the peak to peak amplitude of average signal noise
To be useful, a parameter should provide relevant information and should be
reliable. Onset times have been used for analysing proprioception, but when
studying joint stability, the amount of early activity might be crucial and not
always related to the onset time. It is hypothesized that: a. onset time
detection produces very variable results whereas the calculation of IEMG
provides a higher repeatability and b. onset times and IEMG are not
necessarily related to each other. Therefore the aims of this study were: a. to
test the reproducibility of several onset determination algorithms and
compare it to that from the IEMG and b. expose and compare the information
provided by these parameters.
Chapter one
Onset times and joint stabilisation
12
1.2 METHODS
The EMG signals (1000 Hz) from 23 subjects (12 males, 11 females; 13
stable, 10 unstable) were recorded using bipolar, pre-amplified surface EMG
electrodes placed over the belly of six muscles of the lower leg (peroneus
longus, peroneus brevis, soleus, gastrocnemius lateralis, gastrocnemius
medialis and tibialis anterior) with an interelectrode distance of 2 cm. The
grouping criterion was the subjective feeling of ankle/foot stability as
determined by a short anamnestic questionnaire including among others
frequency, kind and consequences of inversion trauma. Stable and unstable
subjects were analysed separately, since stability could affect the results
(Lynch et al., 1996; Vaes et al., 2001). The subjects underwent expected and
unexpected sudden inversion and eversion trials (20°) while standing on a
tilting platform. The subjects had bare feet and were aware of the tilting
direction. Their left leg was full weight bearing, having its longitudinal axis
parallel to the tilt axis of the plate. The tip of the right foot rested on a block to
help maintain balance (Figure 1.1). The subjects were instructed to look
forward to a spot on the wall. The plate was released from behind the
subjects out of their field of view. Each subject performed three trials per
condition. The instant of release was indicated by counting backwards 3, 2,
1, tilt. At those trials randomly chosen to be unexpected the plate was
released at any time during the countdown. The first trial of each condition
was always unexpected. The onset times were determined using an
algorithm with 6 different parameter combinations (Table 1.2). It was defined
as the time between start of the tilt, as determined by an electrogoniometer
(1000 Hz) aligned with the axis of rotation of the tilt plate, and the instant at
which the filtered signal exceeded the given threshold. Two different median
filters, window widths of 13 and 26 data points respectively, were applied to
the rectified EMG signal (Figure 1.2), since filtering may affect the onset
calculation and as shown by Kadaba et al. (1985) also the repeatability of
EMG phasic activity. As the detection threshold has shown to be a major
factor influencing onset time detection (Winter, 1984) we tested three
different thresholds, namely the mean plus 2, 3 or 4 standard deviations of
Chapter one
Onset times and joint stabilisation
13
the raw rectified activity, recorded prior to the tilting. For IEMG calculation the
EMG signals were rectified and smoothed using a second order Butterworth
filter with a cut-off frequency of 10 Hz. The filtered data were normalised as
follows (Arampatzis et al., 2001).
100
=
k
Fk
Nk
Max
EMG
EMG
EMG
Nk
:
normalised EMG-data from k-muscle
EMG
Fk
:
linear envelope EMG-data from k-muscle
Max
k
:
maximal amplitude of the smoothed signal from k-muscle of each
subject during the first lateral tilt
Then the integrals were calculated from start of the tilt and from each of the
six determined onset times to 200 ms after start of the tilt.
The data was split into eight groups resulting from the combination of the
three dichotomic categories: stability, expectation and tilt direction. To verify
the repeatability of the obtained onset times and integrals we first calculated
the ICCs of all parameters for all three trials from each subject in every
condition to asses the linearity of their relationship. Afterwards the Friedman
test (a non parametric test for k dependent samples) was applied to check for
possible differences (p<0.05) within the set of correlated trials (Yang and
Winter, 1983; Kamen and Caldwell, 1996). Finally, to give an idea of the
absolute difference between trials, the root mean square differences (RMS)
were also calculated.
As an estimator of the sources of variability, the variance due to biovariability
and that due to the measurement were estimated as suggested by
Brinckmann et al. (2002). The coefficient of variance (CV) was then
calculated as the square root of the variance divided by the mean.
Chapter one
Onset times and joint stabilisation
14
2
2
total
bio
R
=
2
2
)
1
(
total
t
measuremen
R
-
=
x
CV
=
2
bio
: variance due to biovariability
2
measurement
: variance due to the
measurement
2
total
: variance of the measured
values
R
: Pearson's correlation
coefficient
x
: mean of the parameter
values
Figure 1.1.
Positioning of the subject on the tilt platform. Left foot was full weight bearing
having its longitudinal axis parallel to the tilt axis of the plate. The tip of the right foot rested
on a block to help maintain balance.
Table 1.2.
The onset times were determined using two filters and three thresholds: Mean
plus 2, 3 and 4 standard deviations (SD).
Median filter. Window width 13 points
Median filter. Window width 26 points
ONSET-1 ONSET-2 ONSET-3 ONSET-4 ONSET-5 ONSET-6
at 2 SD
at 3 SD
at 4 SD
at 2 SD
at 3 SD
at 4 SD
Chapter one
Onset times and joint stabilisation
15
Figure 1.2.
Rectified and filtered (13 and 26 points median filter) EMG signal of the
m.
peroneus longus
during a lateral tilt trial
1.3 RESULTS
An ICC above 0.7 was considered to correspond to an acceptable correla-
tion. It is assumed that when no significant differences (p<0.05) between
correlated trials is found, a higher amount of ICCs above 0.7 corresponds to
a better repeatability. As this does not inform about the magnitude of the
deviation, the RMS values are also reported. As onset times and integrals
have different units, to allow a comparison between these parameters, the
RMS values are additionally expressed as a percentage of the corresponding
means.
The Friedman test only revealed significant differences (p<0.05) between
consecutive measures in 2%, 3% and 6% of the cases for onset times
(ONSET-1 TO 6), integrals from onset to 200 ms after tilt (INTEGRAL-1 TO
6) and integrals from tilt to 200 ms after (INT.-TOTAL), respectively. Since
Chapter one
Onset times and joint stabilisation
16
the results are focusing on repeatability, when citing amounts or percentages
of ICCs above certain level, those trials displaying differences (p<0.05)
between consecutive trials are not included. In order to provide some more
information, in table 1.3 the results are split into more levels of correlation.
Onset times vs. integrals: For ONSET-1 TO 6, thirty six percent of all cases
have an ICC above 0.7. For INTEGRAL-1 TO 6 and for INT.-TOTAL the
corresponding percentages are considerably higher, namely 73% and 81%
respectively. The RMS percentages are also higher for the onsets in almost
all conditions (Tables 1.4 and 1.5). At all other comparisons done below, the
integrals still provide more repeatable results than onset times.
Comparing algorithms: ONSET-1 displays the highest amount of ICCs above
0.7 (44%) and ONSET-3 the lowest one (27%). INTEGRAL-6 has the highest
number of ICCs above 0.7 (81%), whereas INTEGRAL-1 has the least
(58%).
Comparing trial conditions: Unstable subjects seem to produce more
repeatable onset times than stable subjects. No bigger differences are
evident when looking at the integrals. Between trial conditions no further
systematic differences regarding repeatability can be identified (Tables 1.3
and 1.4).
Table 1.3.
Percentage of cases displaying no differences (
p<
0.05) between correlated trials
and having an ICC above given values for the different trial conditions. Data include all
muscles and algorithms.
St-Ux-In St-Ex-In St-Ux-Ev St-Ex-Ev Un-Ux-In Un-Ex-In Un-Ux-Ev Un-Ex-Ev
Ons. Int. Ons. Int. Ons. Int. Ons. Int. Ons. Int. Ons. Int. Ons. Int. Ons. Int.
>0.7
31 67 25 86 39 78 31 75 19 69 50 72 42 89 53 69
>0.8
8 17 8 42 14 72 19 61 17 58 39 50 22 50 39 47
>0.9
0 0 6 8 3 14 6 33 6 8 19 11 14 14 28 19
Stable (St)/Unstable (Un) Unexpected (Ux)/Expected (Ex) Inversion (In) /Eversion (Ev)
Ons. = Onset times / Int. = Integrals
Table 1.4.
Average root mean square differences (RMS) between trials for the different trial
conditions. Absolute values and percentages of the mean. Data include all muscles and
algorithms.
St-Ux-In St-Ex-In St-Ux-Ev St-Ex-Ev Un-Ux-In Un-Ex-In Un-Ux-Ev Un-Ex-Ev
Onset
[ms]
22 29 24 26 20 25 23 20
Int.
[%s]
2.40 2.49 3.91 2.76 1.87 3.06 1.91 1.92
Onset
[%]
42 79 45 70 43 57 49 62
Int.
[%]
32 34 47 35 32 50 40 40
Stable (St)/Unstable (Un) Unexpected (Ux)/Expected (Ex) - Inversion (In) /Eversion (Ev)
Ons. = Onset times / Int. = Integrals
Chapter one
Onset times and joint stabilisation
17
Comparing muscles: Figure 1.3 should help to better illustrate the
repeatability of the parameters regarding the different muscles. The onset
times for m. tibialis anterior and m. peroneus longus show quite more high
ICC values than those from the other muscles. Interestingly when looking at
the integrals, the m. tibialis anterior displays the lowest amount of high ICCs
together with m. peroneus brevis and m. gastrocnemius medialis. Table 1.5
presents the absolute and the relative RMS values for the different muscles.
Figure 1.3 and table 1.5, besides a lower variability for integrals also indicate
that there might be discrepancies between the variability of onset times and
integrals.
Figure 1.3.
Percentage of onset times and Integrals demonstrating an ICC above 0.7 and no
differences (p<0.05) between trials for all studied muscles: PL (m. peroneus longus), PB (m.
peroneus brevis), SOL (m. soleus), GL (m. gastrocnemius lateralis), GM (m. gastrocnemius
medialis) and TA (m. tibialis anterior). Onset-1 and Integral-6 were chosen since they are
showing highest number of ICCs above 0.7 for the respective parameter. In addition Int.-total
is represented, for it is not dependent on onset times.
Table 1.5.
Average root mean square differences (RMS) between trials for the different
muscles. Absolute values and percentages of the mean. Data include all groups and
algorithms.
PL PB SOL GL GM TA mean
Onset [ms]
23 25 27 27 30 24 26
Int [%s]
1.74 1.83 2.32 2.85 3.92 5.10 2.96
Onset [%]
52 73 51 48 50 73 58
Int [%]
36 32 32 45 59 64 45
PL (m. peroneus longus), PB (m. peroneus brevis), SOL (m. soleus), GL (m. gastrocnemius
lateralis), GM (m. gastrocnemius medialis) and TA (m. tibialis anterior).
Onset = Onset times / Int = Integrals
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
PL
PB
SOL
GL
GM
TA
ONSET-1
INTEGRAL-6
INT.-TOTAL
Details
- Seiten
- Erscheinungsform
- Originalausgabe
- Erscheinungsjahr
- 2005
- ISBN (eBook)
- 9783832488352
- ISBN (Paperback)
- 9783838688350
- DOI
- 10.3239/9783832488352
- Dateigröße
- 1.6 MB
- Sprache
- Englisch
- Institution / Hochschule
- Deutsche Sporthochschule Köln – Medizin- und Naturwissenschaften
- Erscheinungsdatum
- 2005 (Juni)
- Note
- 1,0
- Schlagworte
- biomechanik sprungelenk fuss inversion motor
- Produktsicherheit
- Diplom.de
